DeepECG-Reconstructor/analyze_dataclass.py at master · jingweili87/DeepECG-Reconstructor · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155

import numpy as np
import matplotlib.pyplot as plt
import pathlib
import argparse
from explore_functions.compute_statistics import sort_data_matrix, compute_conditional_probability
from util_functions.load_data_ids import load_dataclass_ids
from util_functions.general import get_parent_folder
from util_functions.diagnosis_hierarchy import diagnosis_hierarchy


def analyze_dataclass(data_class: str, var_size = 30, analyze=False):

    """
    This function makes it possible to compute the conditional probability
    for the elements of a given data_class to be associated with the different
    clinical labels used to describe the dataset

    The statistical results are stored in the folder ./../Data/Analysis/Dataclass/data_class/Data/

    The figures describing the figures are stored in the folder ./../Data/Analysis/Dataclass/data_class/Plot/

    """

    parent_folder = get_parent_folder()

    analysis_data_folder = parent_folder + 'Analysis/Dataclass/' + data_class + '/Data/'
    analysis_plot_folder = parent_folder + 'Analysis/Dataclass/' + data_class + '/Plot/'

    pathlib.Path(analysis_data_folder).mkdir(parents=True, exist_ok=True)
    pathlib.Path(analysis_plot_folder).mkdir(parents=True, exist_ok=True)

    try:

        diagnosis_labels = np.load(analysis_data_folder + 'diagnosis_labels.npy')
        conditional_probability = np.load(analysis_data_folder + 'diagnosis_conditional_prob.npy')

    except:

        analyze = True

    if analyze:

        if data_class not in diagnosis_labels:

            exploration_data_folder = parent_folder + 'Exploration/Data/'

            diagnosis_labels = np.load(exploration_data_folder + 'diagnosis_labels.npy')
            diagnosis_matrix = np.load(exploration_data_folder + 'diagnosis_matrix.npy')

            input_indexes = [np.argwhere(diagnosis_labels == input_data_class)[0][0] for input_data_class in diagnosis_hierarchy[data_class]]

            new_column = np.zeros(len(diagnosis_matrix))

            for index in input_indexes:

                new_column += diagnosis_matrix[:, index]

            new_column[np.where(new_column > 0)] = 1

            diagnosis_matrix = np.append(diagnosis_matrix, new_column.reshape((-1, 1)), axis=1)

            diagnosis_labels = np.append(diagnosis_labels, data_class)

            print('Generating conditional probabilities...')

            print()

            diagnosis_matrix, diagnosis_labels = sort_data_matrix(diagnosis_matrix, diagnosis_labels)
            conditional_probability = compute_conditional_probability(diagnosis_matrix, len(diagnosis_matrix), len(diagnosis_labels))

            print()

        else:

            exploration_process_folder = parent_folder + 'Exploration/Process/'

            diagnosis_labels = np.load(exploration_process_folder + 'diagnosis_labels.npy')
            conditional_probability = np.load(exploration_process_folder + 'diagnosis_conditional_prob.npy')

        np.save(analysis_data_folder + 'diagnosis_labels.npy', diagnosis_labels)
        np.save(analysis_data_folder + 'diagnosis_conditional_prob.npy', conditional_probability)

    var_index = np.argwhere(np.asarray(diagnosis_labels) == data_class)[0][0]
    conditional_probability = conditional_probability[var_index].reshape(len(conditional_probability))
    conditional_probability = np.delete(conditional_probability, var_index)
    diagnosis_labels = np.delete(diagnosis_labels, var_index)

    diagnosis_labels = [diagnosis_labels[idx] for idx in (-conditional_probability).argsort()]

    conditional_probability[::-1].sort()

    variable_num = var_size
    conditional_probability = conditional_probability[:variable_num]
    diagnosis_labels = diagnosis_labels[:variable_num]

    _, ax = plt.subplots()

    ax.bar(np.arange(variable_num), conditional_probability)

    ax.set_xticks(np.arange(variable_num))
    ax.set_xticklabels(diagnosis_labels, rotation=90)

    ax.set_yticks(np.linspace(0,1,11))

    ax.set_ylim([0,1])

    ax.set_ylabel('Conditional probability')
    ax.set_xlabel('Diagnosis')

    ax.grid()

    plt.savefig(analysis_plot_folder + 'conditional_probability.png', format='png', bbox_inches='tight')

    plt.cla()
    plt.clf()

    data_size = len(load_dataclass_ids(parent_folder, data_class))

    _, ax = plt.subplots()

    ax.bar(np.arange(variable_num), conditional_probability * data_size)

    ax.set_xticks(np.arange(variable_num))
    ax.set_xticklabels(diagnosis_labels, rotation=90)

    ax.set_yticks(np.linspace(0,data_size,11))

    ax.set_ylim([0,data_size])

    ax.set_ylabel('Subclass sizes')
    ax.set_xlabel('Diagnosis')

    ax.grid()

    plt.savefig(analysis_plot_folder + 'subclass_sizes.png', format='png', bbox_inches='tight')

    plt.cla()
    plt.clf()


if __name__ == '__main__':

    parser = argparse.ArgumentParser()

    parser.add_argument('-data_class', '--data_class', type=str)
    parser.add_argument('-var_size', '--var_size', type=int, default=30)

    args = vars(parser.parse_args())

    data_class = args['data_class']

    var_size = args['var_size']

    analyze_dataclass(data_class, var_size)